OpenAI GPT–6 Astra breaks Enigma message that has resisted solution since 2005
Source Entity
Hacker News
OpenAI's GPT-6 Astra has successfully decrypted a long-standing Enigma code from 1941 that remained unsolved since 2005. Simultaneously, OpenAI and Anthropic have launched more cost-effective AI models, GPT-6 Sol/Luna and Claude Opus 5.5, to remain competitive in the evolving market.
The Breakthrough: GPT-6 Astra and the Enigma Decryption
On September 15, 2026, a significant milestone in computational cryptanalysis was reached when OpenAI’s GPT-6 Astra successfully decrypted the German Army Enigma message MVUEH, dated July 10, 1941. This specific message, transmitted by tactical callsign 2ny and received by the SS-Totenkopf Quartiermeister, had remained an enigma to researchers since 2005. The success of Astra marks a departure from previous human-led efforts, which had only partially succeeded in 2017 when Alex Shovkoplyas decrypted a related message, Nr. 173 (SIPVX), though that effort encountered discrepancies with established daily keys.
Contextualizing the 1941 Enigma Puzzle
The Enigma machine, a centerpiece of World War II communications, relied on complex rotor settings that created a vast number of potential permutations. The MVUEH message has long been a subject of interest for code-breaking enthusiasts because it represents a lingering gap in the historical record of German military communications. The fact that Astra could resolve a sequence that resisted decryption for two decades highlights the leap in pattern recognition and inferential logic capabilities inherent in the GPT-6 architecture.
The Strategic Shift in AI Model Deployment
Beyond historical breakthroughs, the AI industry is currently undergoing a structural shift toward efficiency. OpenAI has introduced GPT-6 Sol and GPT-6 Luna, expanding the GPT-6 family to address specific enterprise needs. Sol is optimized for high-complexity tasks such as advanced coding, while Luna is engineered for high-volume data extraction and summarization. This diversification is a direct response to the market demand for models that balance frontier-level intelligence with operational cost-effectiveness.
Competitive Pressures and Market Dynamics
This move by OpenAI coincides with Anthropic’s release of Claude Opus 5.5, which focuses on token efficiency and a 40% reduction in operating costs. Both companies are navigating a landscape increasingly crowded by open-weight models that provide high utility at lower price points. By slashing API prices by 50% compared to previous GPT-5.6 promotions, OpenAI is signaling a transition from the era of 'experimental' AI to an era of 'utility-based' AI, where cost-per-inference is the primary metric for corporate adoption.
Future Implications for Computational History
The ability of frontier models like Astra to solve historical cryptographic puzzles suggests that AI will become an essential tool for historians and archivists. As these models become more efficient and accessible through tiers like Sol and Luna, the barrier to entry for analyzing complex, unstructured historical data will drop significantly. We are likely to see a wave of similar 'unsolved' historical data being processed, analyzed, and finally understood through the lens of modern large language models, effectively rewriting or filling in the gaps of 20th-century history.